传感器网络拥塞避免与控制的模糊AQM算法

传感器网络节点通信能力有限，当数据到达速率持续超过节点转发能力时网络会发生拥塞；传感器网络是任务型网络，对不同优先级的信息具有不同的服务质量要求.针对传感器网络信息传输的上述特性，提出了一种新的拥塞避免与控制算法FAQM（Fuzzy Active Queue Management）.该算法在综合考虑数据包的随机指数标记概率和优先级权值的基础上，建立了模糊逻辑推理系统，并以数据包丢弃因子作为参量来实现数据流的智能调控.NS2仿真实验结果表明：FAQM算法能减少高优先级数据包的丢弃率和节点间链路的时延，稳定节点队列长度，在有效避免与控制拥塞网络的同时提升网络整体QoS（Quality of Service）性能.

Fuzzy AQM for Congestion Avoidance and Control in Sensor Networks

There is limited communication ability in sensor networks.When the data arrival rate is persistently beyond its transmission ability,the network congestion occurs.Sensor networks are task-based,where data packets with different prior privileges require different Quality of Service (QoS).To address this challenge,an algorithm for congestion avoidance and control in sensor networks,called fuzzy active queue management(FAQM),is proposed.In the algorithm,the random exponential marking probability of a packet along with its priority weight is considered to establish a fuzzy logic inference system,and the intelligent regulation of the data stream is realized by taking the packet dropping factor as the parameter.The experimental results in network simulator version 2 show that FAQM can reduce high priority packets' dropping rate and link delay and stabilize the queue length of the node,therefore it can avoid and control the network congestion and improve the network QoS performance at the same time.